fNIRS-based early identification of mild cognitive impairment - a large-scale multi-paradigm study with ensemble machine learning models

Aug 5, 2026

Upload by Ophelia

Early and accurate identification of mild cognitive impairment (MCI) is essential to enable timely intervention and slow subsequent cognitive decline. Functional near-infrared spectroscopy (fNIRS) offers non-invasive, portable screening, yet performance is limited by single-paradigm feature sets and small samples. This study integrates large-sample resting-state and 1-back task fNIRS data and trains an ensemble machine-learning model to improve diagnostic accuracy and reliability for early MCI screening.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages12

Reading Time30 min

AccessFree to download

Background

Importance of early MCI identification

Role and limitations of fNIRS

Methods

Participants and diagnostic criteria

fNIRS acquisition and preprocessing

Feature extraction and ensemble modeling

Results

Selected features and model performance

Integrated vs single-paradigm comparisons

Discussion

Clinical value and implications

TYPE Original Research PUBLISHED 03 March 2026 DOI 10.3389/fneur.2026.1738099 OPEN ACCESS EDITED BY Mohammad Mofatteh, Queen's University Belfast, United Kingdom REVIEWED BY Florencia Deschle, Hospital Británico de Buenos Aires, Argentina Jin Hyuck Park, Soonchunhyang University, Republic of Korea *CORRESPONDENCE Wenguang Xia [docxwg@163.com](docxwg@163.com) †These authors share first authorship RECEIVED 03 November 2025 REVISED 26 January 2026 ACCEPTED 03 February 2026 PUBLISHED 03 February 2026 CITATION Chong Y, Duan C, Xu X, Li Z, Zhang H, Gong J, Wu Q, Xia L, Zhang P and Xia W (2026) fNIRS-based early identification of mild cognitive impairment: a large-scale multi-paradigm study with ensemble machine learning models. Front. Neurol. 17:1738099. doi: 10.3389/fneur.2026.1738099 COPYRIGHT © 2026 Chong, Duan, Xu, Li, Zhang, Gong, Wu, Xia, Zhang and Xia. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. fNIRS-based early identification of mild cognitive impairment: a large-scale multi-paradigm study with ensemble machine learning models Yufei Chong 1 , 2, 3†, Can Duan3,4, 5†, Xinzi Xu 1 , 2, Zhengliang Li3,4, 5, Heling Zhang 1 , 2, 3, Jingyi Gong 1 , 2, Qingqing Wu3, Lirong Xia 2, Peiwen Zhang 2 and Wenguang Xia 1 , 2, 3,4, 5* 1The Affiliated Hospital of Hubei Provincial Government/Hubei Rehabilitation Hospital, Wuhan, China, 2Hubei Engineering Research Center of Neuromodulation technology, Wuhan, China, 3Hubei University of Chinese medicine, Wuhan, China, 4Hubei Provincial Hospital of Integrated Traditional Chinese and Western Medicine, Wuhan, China, 5Hubei Provincial Clinical Research Center for Stroke Rehabilitation of Integrated Traditional Chinese and Western Medicine, Wuhan, China Background: Early and accurate identification of mild cognitive impairment (MCI) is crucial for timely intervention and preventing further cognitive decline. Functional near-infrared spectroscopy (fNIRS) is a non-invasive, portable tool for clinical screening, but its diagnostic accuracy is often constrained by single-paradigm approaches and small sample sizes. To address this limitation, this study aimed to develop and validate an efficient early MCI screening model by integrating largesample fNIRS data from resting-state and 1-back task paradigms using ensemble machine learning, thereby enhancing the accuracy and reliability of early MCI diagnosis. Methods: A total of 462 right-handed participants (185 MCI patients and 277 healthy controls, aged 58-87 years) were included in the final analysis after screening, with MCI diagnosis jointly determined by two experienced neurologists based on Petersen’s criteria. fNIRS signals were collected during restingstate and 1-back task sessions; after preprocessing in MATLAB, features were extracted from oxygenated hemoglobin (HbO) signals of both paradigms. Results: Feature selection was performed via a gradient boosting classifier based on feature importance scores, resulting in 108 selected features. Five classifiers were trained and evaluated using 10-fold cross-validation. The integrated dataset combining resting-state and 1-back task features outperformed the single-paradigm datasets: the Neural Network model on this integrated dataset achieved an accuracy of 86.49%, sensitivity of 94.74%, specificity of 77.78%, and Area Under the Curve (AUC) of 93.49% . In contrast, the Nearest Neighbor model on the resting-state dataset and the Decision Tree model on the 1-back task dataset yielded accuracies of 70. 27% and 75.68%, respectively. Group classification u

FAQs

What is the main goal of the study on early MCI identification?

To develop and validate an efficient early MCI screening model using large-sample fNIRS data by integrating resting-state and 1-back task paradigms with ensemble machine learning.

How were participants diagnosed with mild cognitive impairment?

MCI diagnoses were jointly determined by two experienced neurologists using Petersen’s criteria.

Why does combining resting-state and 1-back task data improve performance?

The integrated dataset outperformed single-paradigm datasets, suggesting that multi-paradigm feature integration captures complementary information for classification.

Aug 5, 2026

Upload by Ophelia